RE: PFWG-ISSUE-689: Mis-use of ARIA role="heading" needs to be addresed [ARIA 1.1]

Hi,
This is actually something I brought up some weeks back on the UAIG call, and doesn’t refer to the use of role=heading on H# tags, but rather, whether headings must have an implicit heading level in the accessibility tree when aria-level isn’t added in the markup, and whether aria-level should be a required attribute in the authoring spec whenever role=heading is used.

Since there is no such element as <h> my heading </h>, there is no equivalent for a role=heading element that does not include a level, and doing so breaks the outline of the structured order of the page when a heading exists with no level.

So, the proposal is that user agents should set an implicit level when role=heading is used with no aria-level present in the same markup, and the authoring spec should state that aria-level is a required attribute for the use of role=heading to ensure proper outline structuring on the page where its being applied.



From: ahby@aptest.com [mailto:ahby@aptest.com] On Behalf Of Shane McCarron
Sent: Wednesday, December 10, 2014 9:15 AM
To: Protocols and Formats Working Group
Subject: Re: PFWG-ISSUE-689: Mis-use of ARIA role="heading" needs to be addresed [ARIA 1.1]

I disagree that this is a mis-use of heading.  The @role value of heading is correct when applied to headings (e.g., H2 elements).

On Wed, Dec 10, 2014 at 10:01 AM, Protocols and Formats Working Group Issue Tracker <sysbot+tracker@w3.org<mailto:sysbot+tracker@w3.org>> wrote:
PFWG-ISSUE-689: Mis-use of ARIA role="heading" needs to be addresed [ARIA 1.1]

https://www.w3.org/WAI/PF/Group/track/issues/689


Raised by: Richard Schwerdtfeger
On product: ARIA 1.1

Mis-use of the heading role is creating problems for screen readers:
https://github.com/w3c/respec/issues/370


Suggest:

The header role must require the aria-lavel attribute.
1. if heading is used should we require aria-level to be used. ... make it a required attribute
2. We should include a note that says: "if the host language has a an equivalent host language semantic of "heading" it should be used"?





--
Shane McCarron
Managing Director, Applied Testing and Technology, Inc.